I’ll have to go back to using a physical telephone, instead of a headset.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Apple to Phase Out Rosetta 2 Starting With macOS 28 as Intel Era Ends
- Thread starter MacRumors
- Start date
- Sort by reaction score
That's my understanding. Apple are being vague here, but I read it as new app store submissions for macOS 28 will have to be ARM64 code. Older X86 games will still work (for an undefined period of time).
If you needed that in 2019, you bought it to use it. Computers are tools. Not long-term investments. I dropped a mint on an iMac Pro in 2018. I have no regrets about putting it aside in 2020 because the basic M1 MacBook Pro bested it in every metric I relied on for work.
It’s a shame this won’t be repositioned as an SDK. I’ve got enough legacy Intel apps that may or may not be ported that this makes me sad. Rosetta 1 was really messy, but Rosetta 2 feels seamless.
No. Intel MacOS stops at 26. What they are talking about are the security updates to 26. During 26's first year in deployment it is extremely likely it will get security updates that year (along with feature that get updates and bug fixes). In 2027 and 2028 there will be security updates (with likely diminished coverage) also. Egregious security lapse could get fixed in 2029 cycle (but by fall '29 probably done). [ Some new security problems will only appear in macOS 27 and 29 along with the new features that opened the security lapses. ] Very likely no secuirty update that requires a back port of something new will happen. The shared code subsets across macOS major versions that have bug fix that closes a security hole will get a back port.
The 'countdown' clock on the security fixes starts now. ( technically a version 26.0 hasn't even made it out the door yet. )
They dropped support for earlier non-32-bit-clean 68000 apps a long time ago. Maybe around the time of MacOS 7.x?
The announcement about Intel being dropped in MacOS 27 was completely uncalled for, but this one is fine. Developers need to know.
kexts are already deprecated. kext will likely be just "full on dead" once Intel macOS disappears. Rosetta never did cover them.
Apple has a new API called "System Extensions" but they work substantively different than kernel extension. Apple is basically kicking everyone else out of the kernel level. System Extensions works in a 'in between" zone where access is very carefully regulated out in a far more limited fashion.
Especially the code that folks used to create 'hackery' to jump around kernel settings and kernel limitations that Apple put into their code. All the 'raw hardware' , booting loopholes that EFI/UEFI brought along will be gone.
Wind’s howling. Better get one last play through of Witcher 3 via Porting Kit just in case.
This is coming a little earlier than expected. This will again kill gaming on Macs as Apple has done so many times over the last 2 decades.
Once some x86 app has seen that Rosetta 2 is there , some will behave badly if it disappers. Had a upgrade that borked Rosetta for some reason. Had Norton installed. Their firewall drivers were leaning on Rosetta. When emulated borked that startup it borked networking on the machine. Had to work around and get rid of that x86 apps also. ( in catch 22 because can't update to get Rosetta if no networking.)
Pragmatically need to sweep and remove all the Intel binary apps also. That wouldn't been an 'uninstall' as much as a 'garbage collection' on a major OS upgrade. I suspect Apple will just leave things 'messy' though. There is a diminished Rosetta they intend to support incrementally longer. I suspect the smaller subsystem will just 'crash' the more general apps.
Long term though they'll need a garabage collector.
I have yet to find a company that keeps their computers for more than 5 years... (if even that long)
especially if they are willing to spend that much.
While valve has the capability of getting steam built for ARM64, many games that are currently supported on MacOS will completely break, right? I just tried Baldur's Gate 3 and it apears as "Apple" in the Kind category, therefore I'm under the impression that this game will continue to function. For indie developers that don't compile for ARM64, or have abandoned updating their software, is this a death knell for them?
A quote from the end of the new notice.
".. Beyond this timeframe, we will keep a subset of Rosetta functionality aimed at supporting older unmaintained gaming titles, that rely on Intel-based frameworks. .."
About the Rosetta translation environment | Apple Developer Documentation
Learn how Rosetta translates executables, and understand what Rosetta can’t translate.
This probably will eventually get killed off in the extended long term, but it isn't a short term stab. With Metal 4 and probably Metal 5 in 3-4 years at some point better games need to move onto a modern Apple GPU game engine. There is likely are narrow set of games that they'll scaffold over the intermediate term, but at some point gaming industry will have to come up with better porting tools/frameworks.
Growth in mac Games market will over time get more foundational game engines built on Apple GPU specifics and newer version of metal. New games will ported and/or created on that new foundation.
Why drop support for something that works fine and will continue to work fine? They did this with the PowerPC->Intel crossover too leaving so much software useless.
part of OCLP is mucking around with the EFI/UEFI boot process. There is no EFI/UEFI boot process on Apple Silicon macOS. The 'arm only' code doesn't boot the same way. Furthermore, kernel extension that muck around with the boot process... those are also going to go away when Intel dies off also (already deprecated for a while now).
Substantive remaining parts of OCLP is just repackaging/configuring code that Apple (and other kernel developers) already wrote for Intel(x86). There isn't going to be any new stuff.
Kernel security on arm is different.
Gets abandoned. If the developers haven't done a universal binary or moved to a Arm variant after 4 years ( transition kits started in 2020 ) then it has already been abandonded. It isn't about 'abandoned', this is for more so about 'denial'. User denial that their developer hasn't already abandoned the software. Or developer denial that they have a practical business model. Or both.
There are a set of folks who take the "if it ain't broke don't fix it" mentality way too far. It is one where they ignore problems/issues until they break out into a massive 'forest fire' level emergency. If you developer stopped doing updates then you have a 'problem' in a context where the whole platform is going through a migration. Platform moving. App not moving. That is an issue. Burying your head in the sand isn't going to make the issue go away.
It shouldn’t since Windows on ARM has its own translation layer. It isn’t using Rosetta 2.
Windows has its own x86->Arm emulator/compiler .
Apple extended Rosetta 2 to work for Linux x86-64 apps running in a virtual Linux image. ('magic' call outs to the underlying macOS to get some new binaries done).
KB Parallels: Using Rosetta to run x86-64 Docker Containers and Binaries in Linux virtual machines with Parallels Desktop
: the feature is available in Parallels Desktop Pro or Business Edition only.
kb.parallels.com
But those are not 'Windows Apps". Arm on Linux doesn't have a growth problem (steadily getting bigger). Linux server app developers that are not doing Arm builds are largely asleep at the wheel. ( a high fraction of AWS is on Arm now. Same with other server service in the cloud vendors like Oracle/Google/Azure/ etc. ) Likely in part this was Apple eating their own dogfood as they transitioned there server services infrastructure over to Arm machines. Pretty good chance they are about 'done' with that over next 2-3 years.
Crossover leverages Rosetta 2 since they are not really using the underlying Windows foundational subsystems to run x86 wins apps.